Image Merge

The Image Merge (IMRG) was so named because it was the first product that allowed IBM's High Performance Transaction System  (HPTS) to use any of the available conventional reject repair products.  Also,  because image rejects do not have amounts and process control fields encoded, very sophisticated item matching algorithms had to be developed.  These new algorithms greatly improved the merging process for conventional work too.

The Image Merge is a collection of software modules designed to permit the use of any popular reject reentry system when running under (or interfacing to) IBM CPCS. The product is so named because it fills a major gap in the design of the IBM High Performance Transaction System (HPTS) image capture product. It also provides benefits in for integration of third party reject systems into the CPCS environment.

HPTS does not provide an acceptable process for the handling of reject items. HPTS only supports Online Reject Reentry (OLRR). OLRR is a 3270 application that allows codeline corrections to be performed using the paper and the codeline data captured on the sorter. However, OLRR also does not offer any method to correct the physical paper.

This product, Image Merge, provides a way to use a variety of reject repair solutions and integrate their data flows into CPCS and HPTS. This merge is also an evolution of the software technology applied to reject item processing. In addition to offering the key interfaces that allow data flows into HPTS, it also improves the degree of accuracy in reject merge processing. This benefits the operation areas in reduced balancing time requirements, research time and adjustment preparation.

Last, and certainly not least, this merge is designed to handle the item processing requirements associated with unencoded prime capture and merging of repaired rejects into the data stream. Unencoded prime capture runs do not have the amount (and in many cases a process control field) encoded on the check. This reduces the available fields (and characters) for a reject data merge match by 33-50%. Image Merge has unique multi-phase algorithms that determine reject item locations in the prime data stream by matching the higher probability items first. There are eight phases employed, each building on the results of the prior, more certain, merge phases.

The software automatically determines the data flow type (image or conventional check) and selects the appropriate input strings for processing.  These flows are illustrated in Conventional Flow and Image Flow.

The software permits the merge processing to be automated. This feature eliminates a manual step in the operation.

OLRR Bypass is required in Image Processing using HPTS. This task removes a manual step in the operation to accelerate the begin-time for Key Entry.

The repaired string of data may be an R-String or an I-String generated by one of the more popular reject reentry systems. This string can contain the rejects from many different prime pass runs. The prime runs may be image or conventional and can be intermixed.

The Image Merge interfaces with user exit routines (optional) that can be used to deal with unusual data flows. The exit routines have the ability to "force" a match of two items or to deny a match of two matched items.

IMRG produces merge statistics and control information that may be used by existing merge analysis reporting programs or fed into a user-written reporting module. These statistics provide an absolute audit trail of why each reject item merged where it did. They also provide a cross-reference of the reject system assigned item sequence numbers and the prime pass sequence numbers.

Other CPCS tasks can be autostarted by IMRG processing through user customized task start lists.

Accuracy of merge processing has been greatly improved by the addition of several new item location assignment algorithms and improvements on the matching process.

The same executable IMRG software suite can process both subset and non-subset entries. This allows increased controls for conversion and more efficient workflows.

IMRG merges a prime pass I-String, HSRR I-String and an R-String (OLRR to HSRR). For reject systems other than RPA, a change to the R-String is required to enable this feature.

The IMAGE Merge task replaces the CPCS MRGE function for corrected rejects. This merge is different from the PMRG/BMRG tasks (Lundy) currently used in the following ways:

1. It runs under CPCS using string input.

2. It contains additional codeline matching algorithms to improve merge efficiency.

3. It contains provision to use RECO data for the codeline so that when such information is available, the merge can approach perfection in the match results.

4. It provides for a later level of software that uses an on-line, image-based, operator assist merge function.

5. The merge occurs after the HSRR I-String or R-String is created. This means kill lists will match the order of items in the pockets unlike the output of the PMRG/BMRG tasks.

 

 

Send mail to Webmaster@TranSoftware.com with questions or comments about this web site.
Copyright © 1999 Transaction Sofware, Inc.
Last modified: March 03, 2008 10:52